Content uploaded by Mohammad Rahim Rahnama
Author content
All content in this area was uploaded by Mohammad Rahim Rahnama on Aug 02, 2023
Content may be subject to copyright.
Applied Geography 124 (2020) 102318
0143-6228/© 2020 Published by Elsevier Ltd.
A spatial-temporal analysis of urban growth in melbourne; Were local
government areas moving toward compact or sprawl from 2001–2016?
Mohammad Rahim Rahnama
a
,
*
, Ray Wyatt
b
, Lia Shaddel
a
a
Geography and Urban Planning, Ferdowsi University of Mashhad, Mashhad, Iran
b
Urban Planning, Deakin University, Australia
ARTICLE INFO
Keywords:
Compact
Melbourne
Spatial-temporal analysis
Sprawl
Urban growth
ABSTRACT
In most cities, urban growth follows a sprawl pattern. In the Melbourne metropolitan area, the Plan Melbourne
(2017–2050) was formulated with the aim of reducing sprawl and encouraging sustainable growth. However, the
Internet research has shown that in the metropolitan area of Melbourne, no studies aiming to investigate the
urban growth using satellite images and analyzing compact urban growth drivers (CUGDs) have been conducted.
The objectives of this study of the Melbourne metropolitan area are as follows: 1) Analyzing spatial-temporal
changes of normalized difference built-up index (NDBI) 2) Analyzing CUGDs 3) Evaluating the relationship
between NDBI and CUGDs. 4) Determining the type of growth in metropolitan area divisions regarding NDBI and
CUGDs (2001–2016). To calculate NDBI and urban built-up areas Landsat satellite images were used. Five in-
dicators of population density, separate house density, apartment house density, public transportation usage
ratio and distance from the city center were analyzed as CUGDs. The relationship between NDBI and CUGDs was
assessed using Ordinary Least Squares (OLS). GIS and ENVI software packages were used for these analyses. The
ndings revealed that outer Melbourne has experienced the highest rate of change in built-up areas and the
direction of developments are toward west, north and south-east in conformity with the government policies. The
highest annual rate of change belongs to apartment house density. During 2001–2006, it was 67%, during
2006–2011 45%, during 2011–2016 58%, and during 2001–2016, 61% of the variation in NDBI was explained by
the variables. Changes in NDBI under various CUGDs conditions led to the formation of patterns conceptualized
at the level of thirty-one local governments (LGs). Based on this study’s ndings, practical strategies should be
formulated to guide the future development of the city and to achieve the Plan Melbourne (2017–2050)
objective.
1. Introduction
The world urban population has doubled in 43 years, but urban land
cover has doubled in 19 years (Angel, Parent, Civco, Blei, & Potere,
2011) as cities have emerged as economic drivers and dynamic spaces
(Mendiratta & Gedam, 2018; Shirowzhan & Sepasgozar, 2019). Urban
growth is a global phenomenon and is the most irreversible type of
urban land change (Calzada, Meave, Bonl, & Figueroa., 2018; Sharaf,
Serra, & Saurí, 2018). It seems that urban growth in most parts of the
world follows a sprawl pattern (Wolff, Haase, & Haase, 2018).
Sprawl growth is a dynamic spatial phenomenon dened as scattered
and ineffective growth (Kumar, 2017). Sprawl development suffers from
complexity and difculties in denition. Some researchers have recog-
nized sprawl as a multidimensional phenomenon. The development
characterized by widespread, single story low density building and
roadways are the most obvious format of sprawl urban development
(Abdullahi, Pradhan, & Al-sharif, 2017; Ewing, 1997). Various studies
have been conducted on sprawl (Angel et al., 2011; de Espindola, da
Costa Carneiro, & Façanha, 2017; Liu He, Tan, Liu, & Yin, 2016). In
2005, Theobald studied the urban sprawl pattern beyond the urban
boundaries and predicted the residential density of the U.S. in 2020.
In China, after 2000, under the control of the central government for
protecting farmland and food security, less land was available for
dispersed urban development. In response to the shortage of land for
expansion, urban development transformed from dispersed to compact
(Liu, Zhang, Kong, Wang, & Chen, 2018). In some instances, urban
sprawl is planned, for example the urban development of Al-Ain took
place in accordance with its two master plans and vast residential and
* Corresponding author.
E-mail addresses: rahnama@um.ac.ir (M.R. Rahnama), lia.shaddel@gmail.com (L. Shaddel).
Contents lists available at ScienceDirect
Applied Geography
journal homepage: http://www.elsevier.com/locate/apgeog
https://doi.org/10.1016/j.apgeog.2020.102318
Received 23 July 2019; Received in revised form 16 August 2020; Accepted 21 August 2020
Applied Geography 124 (2020) 102318
2
service land uses were added to the city (Sharaf et al., 2018).
This type of growth has negative economic, social and ecological
consequences (Henriquez, Az´
ocar, & Romero, 2006; Lal Kumar, &
Kumar, 2017; Ogle, Delparte, & Sanger, 2017). For example, its effect on
air pollution (Mendiratta & Gedam, 2018), increased urban heat islands
(Ogle et al., 2017; Pearsall, 2017), energy use (Güneralp et al., 2017),
and urban slums (Kohli Sliuzas, & Stein, 2016).
In response to these challenges, compact urban growth was proposed
(Haase, Kabisch, & Haase, 2013). This improves livability and has many
other advantages (Kotharkar, Bahadure, & Sarda, 2014; Mouratidis,
2017). Compact urban growth policies have positive effects not only on
metropolitan scale, but also on a neighborhood unit level (Lim & Kain,
2016). However, these policies are debatable; for instance, it is argued
that increased density might have negative effects on resident satisfac-
tion (Bramley, Dempsey, Power, Brown, & Watkins, 2009), air pollution
(Schindler & Caruso, 2014), or trafc jams (Güneralp et al., 2017).
Nevertheless, these issues can be mitigated through urban planning and
design (Lim & Kain, 2016). Overall, studies show that a compact city is a
type of sustainable urban growth (Kotharkar et al., 2014; Lee et al.,
2014).
Consumption of urban land is in contradiction with the concept of
compact city (Westerink et al., 2013). The NDBI is one of the indices of
sprawl or compactness of urban growth which can be effectively used to
identify urban built-up features and their areas and provide a valuable
view of changes in urban land cover (Grifths, Hostert, Gruebner, & van
der Linden, 2010; Naserikia, Asadi Shamsabadi, Raeian, & Leal Filho,
2019). To better understand urban growth and to guide it toward
compactness, it is important to identify CUGDs (Li, Li, & Wu, 2018).
Some simple but powerful indicators for describing compact urban
growth exist, such as population density (Betru, Tolera, Sahle, & Kassa,
2019; Burton, 2002; Wolff, Haase, Haase, & Kabisch, 2016), public
transportation usage ratio (Burton, 2002; Lee et al., 2014; Liao & Wei,
2014; Mouratidis, 2017), walking and biking levels in urban environ-
ments (OECD, 2012), types of housing (Haase et al., 2013; Shirowzhan &
Sepasgozar, 2019; Shum & Watanabe, 2017) and distance from the city
center (Achmad, Hasyim, Dahlan, & Aulia, 2015; Angell et al., 2011).
Although changes of NDBI in different periods and city areas can be
observed, it is useful to study its relationship with the aforementioned
drivers as they can lead to different urban growth spatial patterns.
Various studies have used remote sensing data as the primary
method to investigate aspects of urban growth and measure urban
sprawl (de Espindola et al., 2017; Liu et al., 2018; Mendiratta & Gedam,
2018). Along with satellite images, GIS can prove useful in under-
standing changes (Achmad et al., 2015).
Urban sprawl and suburbia are the most prevalent forms of urban
growth in Australia, and Melbourne is a leader in both. In 2001, the
Melbourne 2030 vision was drafted with the aim of managing city
growth. In this vision, planning for compact urban growth was
emphasized to the extent that” a more compact city” was its rst
objective. It seems that in this plan, the drivers which affect the growth
of Melbourne have not been identied correctly. Since population grew
faster than predicted in the plan, in 2017 the Plan Melbourne
(2017–2050) was developed. This emphasis on that Melbourne grows
sustainably, and urban sprawl is reduced. However, the Internet
research has shown that in the metropolitan area of Melbourne, no
studies aiming to investigate the urban growth using satellite images and
analyzing compact urban growth drivers (CUGDs) has been conducted.
If urban form is an outcome of the urban growth process, guiding
growth to produce a favorable outcome is important. By identifying
main drivers and analyzing them over time, urban planners can do this
(Bhatta, Saraswati, & Bandyopadhyay, 2010). Therefore, the objectives
of this study on the metropolitan area of Melbourne are as follows: 1)
analyzing spatial-temporal changes of NDBI, 2) analyzing CUGDs at the
level of 31 LGs, 3) evaluating the relationship between NDBI and
CUGDs, 4) identifying urban growth type in metropolitan area divisions
with regards to NDBI and CUGDs (2001–2016).
2. Study area
Melbourne is the capital of the Australian state of Victoria. The
metropolitan area is comprised of 31 LGs (Fig. 1). Melbourne is the
second city in Australia and is a leader in terms of economy, tourism,
and services. According to the latest census conducted in Australia in
2016, the metropolitan area spans 8833.1 square kilometers and has a
population of 4,415,382.
In 2001, the Melbourne 2030 plan was formulated for the metro-
politan area. According to this, population was forecast to reach 4.7
million and a sustainable environment with less sprawl should be
created. In this plan, a boundary for urban growth was created with the
aim of guiding growth to areas capable of offering proper infrastructure
and services and protecting valuable suburban land. However, a greater
than expected growth in population led to the formulation of the Plan
Melbourne (2017–2050) and the boundary for urban growth was
revised. According to the new plan, population was forecast to reach 8
million and 1.6 new dwellings should be created. The plan directions
included, exploiting land effectively, creating an integrated trans-
portation system, providing medium and high-density housing settle-
ments close to service centers, protecting fertile agricultural lands,
encouraging walking and biking, and reducing sprawl growth. Thus,
compact urban growth was emphasized in this plan.
Considering the Melbourne plans for 2030 and 2050 this study’s
analyses were conducted in the metropolitan area. Generally, studies of
urban growth concentrate on major cities and metropolitan areas
(Paulsen, 2014).
3. Method
The analysis was conducted for the periods 2001, 2006, 2011, and
2016. The Plan Melbourne (2017–2050), which started in 2017, was the
main reason for choosing this time span and this study will identify
major changes for achieving the plan objectives and formulating effec-
tive strategies.
3.1. Data
3.1.1. NDBI
Zha and Colleagues (2003) presented NDBI for mapping built-up
areas. The value of this index ranges from −1 to +1 for each pixel. A
positive number is indicative of build-up and is indicative of density
levels. NDBI is calculated using eq. (1) (Naserikia et al., 2019) where bni
and bmi are the digital numbers of mid-infrared and near-infrared bands
of the Landsat images respectively.
NDBI =(bni −bmi)
(bni +bmi’)(1)
Landsat satellite images for 2001, 2006, 2011, and 2016 were
downloaded from the USGS website (Data Set: Landsat 7 ETM +C1
Level −1) and are located on path 92, row 86, 87 and path 93, row 86,
87. To calculate this index, bands 4 and 5 were used. All these satellite
images were acquired in November, so the impact of seasonal variations
was insignicant. The images were geometrically corrected with the
standard projection system (UTM 55 S, WGS1984 datum).
3.1.2. CUGDs
Many indicators exist for evaluating compactness and several driving
factors are usually considered for urban growth analysis (Zhang et al.,
2013). Those used in this study were as follows:
1: Population Density: increased population density over time is
considered as one of the drivers for compact urban growth (Angell
et al., 2011; Kotharkar et al., 2014; Liu et al., 2018; Wolff et al.,
2016). According to these studies, to calculate this, population is
divided by surface area. However, population density on its own
M.R. Rahnama et al.
Applied Geography 124 (2020) 102318
3
cannot be indicative of urban sprawl or compactness. Therefore, four
drivers were also analyzed.
2: Separate House Density; 3: Apartment House Density: A 10-
storey building can provide housing for 10 families, obviating the
need for each family to settle in a separate house. Therefore, type of
housing is very important in determining the type of urban growth
(Shirowzhan & Sepasgozar, 2019; Shum & Watanabe, 2017). This
driver is calculated by dividing separate or apartment house surface
area by total surface area. However, in this study, due to unavail-
ability of the data regarding the surface area of various types of
housing, a number of separate and apartment houses were used.
Therefore, to calculate drivers 2 and 3, the number of separate and
apartment houses were divided by the total surface area.
4: Public Transportation Usage Ratio: Public transportation sys-
tems are regarded as an important tool for the realization of compact
urban growth (Angel et al., 2011; Burton, 2000; Liao & Wei, 2014).
Levels of pedestrian transportation and usage of bicycles are also
another indicator (McCormack & Shiell, 2011; OECD, 2012). To
calculate this driver, the data provided by “method of travel to work
(one, two, and three methods)” was used. The sum of travels made by
public transport, pedestrian, and bicycle was calculated and divided
by the total number of travels.
5: Distance from the City Center: Distance from the city center is a
driving force of urban growth (Achmad et al., 2015; Kotharkar et al.,
2014). Using a “point distance” command in GIS, the distance be-
tween each local government and the city center was calculated.
To calculate the rst four drivers, the required data for local-
government level was downloaded from the Australian Bureau of Sta-
tistics (https://www.abs.gov.au) for 2001, 2005, 2011 and 2016.
3.2. Analysis method
The downloaded images were mosaic after geometric correction and
then clipped. Eq. (1) was calculated using bands 4 and 5 for 2001, 2006,
2011, and 2016, so that spatial-temporal changes were revealed. In
addition, images were classied according to supervised classication to
calculate the built-up areas surface. Following the calculation of CUGDs,
a variable change rate (VCR) was drawn. VCR was calculated using eq.
(2) and is shown at the metropolitan area level.
VCR =A2016 −A2001
A2001 ×100 (2)
VCR shows the rate of change in the variable. A2016 and A2001
show the amount of variable in 2016 and 2001.
The relationship between NDBI and CUGDs was assessed using OLS.
This is the most common method for estimating unknown parameters in
a linear regression model where output is shown by a formula (Bun &
Harrison, 2018). GIS and ENVI software packages were used to conduct
the analyses.
To formulate policies to prevent urban sprawl, it is necessary to
provide a typology depicting how changes in variables lead to the for-
mation of various urban growth patterns. Changes of NDBI under
various conditions of CUGDs led to the formation of various patterns
conceptualized in Fig. 2. The First Annual Growth Rate (AGR) should be
calculated to determine the amount of change.
AGR =100% ×⎡
⎢
⎣(Aend
Astart)1
d
−1⎤
⎥
⎦(3)
AGR is the annual growth rate of the variable. A
start
and A
end
show
the quantity of changes at the beginning and end of the period and d is
the time gap between the two periods.
NDBI is compared separately with each driver. If NDBI and separate
house density increase, both situations, with varying growth rates, lead
to the formation of a sprawl pattern. In other words, single family
separate housing develops. If NDBI increases but apartment house
density decreases, this is also indicative of sprawl growth; however, an
increase in apartment house density is a sign of moving toward
compactness. If the rate of NDBI growth is higher than an increase in
apartment house density, a contrast between sprawl and compact pat-
terns is created. A decrease in NDBI and an increase in apartment house
density is a sign of increase in high-rise buildings and compactness.
If the rate of NDBI growth is higher than the population density
growth rate, this is a sign of sprawl development (rst case), but if the
population density growth rate is higher than the rate of NDBI growth, it
is a sign of compact development (second case). A reduction in popu-
lation density owing to increased built-up areas is indicative of sprawl
shrinkage. Increase in population density together with a decrease, or no
increase in built-up areas, is a sign of reusing existing buildings. A
decrease in population density and a decrease in NDBI lead to compact
shrinkage where the rate of decrease in NDBI is higher than population
density reduction rate. Under such circumstances, the position of the LG
in terms of development deteriorates and withers.
An increase in NDBI and a decrease in public transportation usage
ratio is a sign of moving toward sprawl. On the other hand, an increase
in public transportation usage ratio and a decrease in NDBI shows that
the background for compact growth is being prepared.
An increase in NDBI along with an increase in distance from the city
center leads to sprawl and an increase in NDBI along with a decrease in
distance from the city center is a sign of compact growth. The threshold
of distance from the city center was calculated using “standard distance”
in GIS and a distance of 26.6 Kilometers was obtained. Fig. 2 was
conceptualized for 31 LGs of the city of Melbourne, from 2001 to 2016.
Fig. 1. Location of Melbourne and Metropolitan area divisions.
M.R. Rahnama et al.
Applied Geography 124 (2020) 102318
4
4. Results
4.1. Spatial-temporal changes of NDBI
Table 1 shows that built-up areas have increased from 1609.70 km
2
in 2001 to 1984.12 km
2
in 2016 with an AGR of 1.265%.
The VCR of built-up areas over the 15-year period equals 23.26%.
The highest rate is observed in outer Melbourne (Fig. 3). The central
area, owing to insufcient vacant land, has experienced less change
(Sharaf et al., 2018).
Melbourne is Australia’s second city of and is a popular immigration
or travel destination which has led to an increase in built-up areas. The
spatial changes of NDBI can be clearly seen in Fig. 4.
The observed growth is caused by government policies which have
guided the development toward west, north and south-east. Agricultural
lands located in the east of the city are major limitations for growth.
Melbourne has experienced two growth waves and is currently experi-
encing its third. It is important to manage this third growth wave such
that citizens of Melbourne can benet from it.
To evaluate accuracy, the Kappa coefcient was calculated which is
indicative of similarity between ndings and reality (Liu et al., 2016). A
coefcient from 0.6 to 0.8 indicates high accuracy (Zheng et al., 2015).
This coefcient for each of the years 2001, 2006, 2011 and 2016 was
calculated to be 0.91, 0.87, 0.83 and 0.85 respectively.
4.2. Analysing CUGDs
First, the annual growth rate (Eq. (3)) was calculated for each driver.
Given the results listed in Table 2, all drivers had positive growth during
the 15 year period. Apartment house density had the highest annual
growth rate. Next, the variable change rate was calculated using eq. (2)
(2001–2016). The results obtained for each driver were mapped at the
level of local governments (Fig. 5).
The VCR of population density (Fig. 5a): Over the 15-year period,
this driver has been an increasing trend in all LGs and no decrease is
observed. Melton, Wyndham, and Yarra have seen the biggest change in
the population density. With an increase in the distance from the city
center, the population density generally decreases; however, a density
increment is observed in ve local governments: Maroondah, Kingston,
Glen Eira, Stonnington, Port Philip.
The VCR of separate house density (Fig. 5b): The VCR of separate
house density shows that Wyndham and Melton have had the most
positive changes and Port Phillip and Stonnington in central Melbourne
have had the most negative. With an increase in the distance from the
city center, the density of separate house increases; however, a decrease
in the density of separate house is observed in ve local governments:
Port Philip, Stonnington, Glen Eira, Whittlesea, and Mornington
Peninsula.
The VCR of apartment house density (Fig. 5c): The largest positive
change for the apartment house density is seen in Melbourne © and the
largest negative in Yarra Ranges and Cardinia. In the Port Phillip,
Stonnington and Glen Eira locals, as the distance from the city center
increases, the population density increases and the density of separate
house decreases. Thus, it can be said that in these locals, the density of
apartment house increases, which is also observed in Fig. 5c.
The VCR of public transportation usage ratio (Fig. 5d): The VCR of
public transportation usage ratio shows the largest positive change in
Wyndham and the largest negative in Yarra Ranges and Mornington
Peninsula. With more distance from the city center, the ratio of public
transportation usage decrease.
4.3. Evaluating the relationship between NDBI and CUGDs
To evaluate the relationship between NDBI and CUGDs in the Mel-
bourne metropolitan area, OLS was used. Each driver was calculated
using equation (2) and then used as input for the OLS model. Here, Z (the
dependent variable) is a linear combination of independent variables.
The regression equation is as follows:Where b
0
is the intercept of the
model, b
i
=(i =0, 1, 2, …n) represents the regression model coefcients,
and x
i
=(i =0, 1, 2, …, n) represents the independent variable.
According to Table 3, in 2001–2006 67%, in 2006–2011 45%, in
2011–2016 58% and from 2001 to 2016 61% of the variance in NDBI is
explained by the variables of the model. According to this table, the
regression equations for each of the periods are as follows:
2001–2006:
NDBI = − 16.89(POPD) −1.3(AHD) +13.52(SHD)-3.6 (PTUR) +2.66
(DCC) +311.605.
2006–2011:
NDBI = − 12.16(POPD) −2.2(PTUR) +111.42.
2011–2016:
NDBI = − 14.38(POPD) −1.52(AHD) +2.13 (DCC) +157.233.
2001–2016:
NDBI = − 15.73(POPD) −1.83(AHD) +12.76(SHD) −5.2(PTUR)
+3.08(DCC) +274.519.
SHD: Separate House Density.
AHD: Apartment House Density.
POPD: Population Density.
PTUR: Public Transportation Usage Ratio.
DCC: Distance from the City Center.
The drivers of population density, apartment house density and
public transportation usage ratio have a negative relationship with NDBI
Fig. 2. Patterns of urban growth due to changes in
NDBI and CUGDs
SHD: Separate House Density
AHD: Apartment House Density
POPD: Population Density
PTUR: Public Transportation Usage Ratio
DCC: Distance from the City Center.
Table 1
Built-up areas in 2001, 2006, 2011, 2016
Built-up areas AGR VCR (2001–2016)
2001 1609.70 2001–2006 2.22%
2006 1797.09
2011 1868.37 2006–2011 0.77% 23.26%
2016 1984.12 2011–2016 1.2%
M.R. Rahnama et al.
Applied Geography 124 (2020) 102318
5
and separate house density, and distance from the city center has a direct
and positive relationship with NDBI. The signicance of this relationship
in various periods is depicted in Table 2. Population density has a sig-
nicant, negative relationship with NDBI in all four periods. A one-unit
increase in this driver leads to a 16.89 (2001–2006), 12.61
(2006–2011), 14.38(2011–2016) and 15.73 (2001–2016) unit decrease
in the variance of built-up areas in the metropolitan area.
4.4. Identifying the type of urban growth of metropolitan area divisions
Evaluation of the relationship between NDBI and CUGDs (Table 2)
reveals which of the variables have a signicant relationship with the
increase in metropolitan built-up areas. At this stage, Fig. 2 was
conceptualized for 31 LGs during 2001–2016 (Fig. 6) until types of
growth were identied in metropolitan area divisions (in this period, all
the variables have a signicant relationship with the variance of NDBI).
5. Discussion
Melbourne is a city which is highly inuenced by population ow
and immigration. Also, according to the Plan Melbourne (2017–2050),
1.6 million new dwellings should be created. Therefore, controlling and
guiding its spatial structure toward a compact and sustainable city is
very important. However, analyzing the spatial structure of a city is a
Fig. 3. VCR of built-up areas.
Fig. 4. NDBI a) 2001 b) 2006 c) 2011 d) 2016.
Table 2
The quantity of CUGDs.
Population Density AGR Separate House Density AGR Apartment House Density AGR Public Transportation Usage Ratio AGR
2001 15.10 1.2% 4.5 0.5% 1.47 2.07% 0.152 1.37%
2006 15.69 4.4 1.55 0.154
2011 16.79 4.6 1.76 0.175
2016 18.65 4.9 2.07 0.191
M.R. Rahnama et al.
Applied Geography 124 (2020) 102318
6
complex and challenging undertaking since it is the outcome of events,
technologies, priorities, and history. (Kotharkar et al., 2014).
The results of Table 1 showed that the built up area has increased
from 2001 to 2016 and Outer Melbourne has the largest rate of change.
A study of the spatial variations of NDBI (Fig. 4) conrms these ndings
as well. Thus, as land loss is one of the environmental problems of urban
sprawl, the metropolitan area of Melbourne has expanded (de Espindola
et al., 2017). The compact city paradigm protects the land as one of the
national capitals (Shum & Watanabe, 2017).
Generally, urban compactness is achieved using three approaches. In
the rst, governments plan for compactness by guiding development
toward areas close to public transportation. In the second, compactness
is achieved by redeveloping existing residential sites. In the third,
compactness is obtained by increasing building density in urban plans.
Australian state governments usually adopt the second approach, avoid
the rst, but also incline toward the third (Buxton & Tieman, 2005).
Built-up areas in Melbourne have a low annual growth rate (Table 1).
However, over 15 years, population density has increased (Table 2)
indicating the second approach has been adopted to make Melbourne
more compact. This policy is clear in seven LGs of inner Melbourne
(Fig. 6c).
Population density and its spatial distribution are fundamental ele-
ments of urban form (Bettencourt & West, 2010; Mendiratta & Gedam,
2018). In this study this variable has a signicant relationship with the
variance of NDBI in all four periods (Table 2). In fact, urban growth is
inclined to occur in places where population density is high (Achmad
et al., 2015; Haase et al., 2013). As can be seen in Figs. 3–5 (a), the
highest rate of change in built-up areas and population density corre-
spond to one another and are observed in Wyndham, Melton, Whit-
tlesea, Cardinia, and Casey. This ow is also compatible with
government policies for guiding growth toward the west, south-east and
north. However, the highest rate of change in separate house density is
observed in Wyndham and Melton and, due to an increase in NDBI and
distance from the city center, these LGs are moving toward sprawl. In
Fig. 5. VCR of a) Population Density b) Separate House Density c) Apartment House Density d) Public Transportation Usage Ratio.
Table 3
The results of OLS.
2001–2006 2006–2011 2011–2016 2001–2016
coefcient p. value R2 coefcient p. value R2 coefcient p. value R2 coefcient p. value R2
Population Density −16.89 0.001 0.67 −12.61 0.003 0.45 −14.38 0.002 0.58 −15.73 0.001 0.61
Apartment House Density −1.30 0.04 −2.18 0.4 −1.52 0.03 −1.83 0.04
Separate House Density 13.25 0.004 15.8 0.5 10.11 0.6 12.76 0.006
Public Transportation Usage Ratio −3.6 0.02 −2.2 0.04 −4.12 0.4 −5.2 0.03
Distance from the City Center 2.66 0.03 3.76 0.9 2.13 0.03 3.08 0.03
M.R. Rahnama et al.
Applied Geography 124 (2020) 102318
7
Fig. 6. Patterns of urban growth under various NDBI conditions CUGDs (2001–2016) a) Separate House Density (SHD) b) Apartment House Density (AHD) c)
Population Density (POPD) d) Public Transportation Usage Ratio (PTUR) e) Distance from the City Center (DCC).
M.R. Rahnama et al.
Applied Geography 124 (2020) 102318
8
Cardinia, increased NDBI, increased separate house density, low apart-
ment house density, increased population density, decreased levels of
public transportation usage ratio and increased distance from the city
center have caused sprawl. Increased NDBI, decreased apartment house
density, and increased population density are also indicative of sprawl
growth in Casey. Similarly, Whittlesea, considering its increased NDBI,
increased separate house density and increased distance from the city
center, is moving toward sprawl (Figs. 5 and 6). Therefore, the gov-
ernment needs to adopt more effective strategies to guide development
in the mentioned directions and it seems that a more comprehensive and
integrated plan is required.
In addition, from 2001 to 2016, population density has not decreased
in any of the LGs and thus there is no shrinking (Fig. 6c). In shrinking
cities, density is reduced owing to population reduction which leads to
excess supply and inefcient usage of urban lands (Wolff et al., 2018).
Transportation is another factor that affects urban growth (Liao &
Wei, 2014). This driver has a signicant relationship with the variance
of NDBI in two periods (2001–2016, 2006–2011). During the rst wave
of growth in Melbourne, mechanical transportation enabled the city to
develop toward the suburbs and increased ownership of personal vehi-
cles (the second wave) exacerbated this sprawl growth. Rahnama and
colleagues (2015) showed that in Melbourne personal vehicle trans-
portation is more dominant than other forms. However, public trans-
portation usage ratio has had an increasing trend and Wyndham and
inner Melbourne have had the biggest positive changes and Yarra
Ranges and Cardinia the smallest. In other words, as distance from the
city center increases, public transportation usage ratio decreases
(Fig. 5d) which is in contradiction with the concept of compact cities.
The inuence of separate house density and apartment house density
in the variance of NDBI was conrmed in the OLS. This nding can be
interpreted using the ndings of Haase and colleagues (2013), de
Espindola and colleagues (2017) and Angel and colleagues (2011), who
showed that housing type affects urban growth and spatial structure.
During 2001–2016, the AGR of apartment house density has been higher
than separate house density (Table 2). The biggest VCR in the apartment
house density and separate house density has occurred in inner and
outer Melbourne, respectively (Fig. 5). In other words, according to
these two variables, outer LGs have moved toward sprawl and inner LGs
have moved toward compactness (Fig. 6 a, b).
An increase in the distance from the city center has a signicant
relationship with the variance of NDBI. Studies show that distance from
the city center is one of the driving forces affecting urban growth
(Achmad et al., 2015). As the distance from the city center increases, a
sprawl spatial structure is observed in LGs, which is in contradiction
with the theory of compact cities (Rahnama et al., 2015). Therefore,
these analyses can be useful to better understand existing spatial
structure and urban growth over time.
6. Conclusion
To compile the appropriate policies which prevent the urban sprawl,
a growth model is needed that considers the mechanisms of spatial
variations. NDBI and CUGDs were analyzed to study the spatial varia-
tions and determining the type of growth of each local government in
Melbourne, a forerunner in the urban sprawl in Australia. Outer Mel-
bourne has had the greatest changes in built up urban areas. From 2001
to 2016, 61% of NDBI variation has been explained by 5 CUGDs. The
density index of apartment units has the largest annual growth rate.
Considering the land shortage in the downtown area, the variations of
this index in this area have been low.
The annual rate of public transport usage has been 1.37%. Consid-
ering the signicance of public transport in compact urban form,
particular attention should be paid to this issue. Based on Plan Mel-
bourne (2017–2050), Melbourne should move towards sustainability
and compactness. With due attention to the signicance of urban
growth, the growth and urban form could be investigated beyond the
urban boundaries so that the growth integrity be preserved.
Funding
This research received no specic grant from any funding agency in
the public, commercial, or not - for prot sectors.
Declarations of competing interest
None.
CRediT authorship contribution statement
Mohammad Rahim Rahnama: Conceptualization, Methodology,
Software, Formal analysis, Investigation, Data curation, Supervision,
Project administration. Ray Wyatt: Validation, Resources, Writing -
original draft, Writing - review & editing. Lia Shaddel: Conceptualiza-
tion, Methodology, Software, Formal analysis, Investigation, Resources,
Data curation, Visualization.
References
Abdullahi, S., Pradhan, B., & Al-sharif, A. A. (2017). Sprawl versus compact
development. In Spatial modeling and Assessment of urban form (pp. 35–58). Cham:
Springer. https://doi.org/10.1007/978-3-319-54217-1_3.
Achmad, A., Hasyim, S., Dahlan, B., & Aulia, D. N. (2015). Modeling of urban growth in
tsunami-prone city using logistic regression: Analysis of Banda Aceh, Indonesia.
Applied Geography, 62, 237–246. https://doi.org/10.1016/j.apgeog.2015.05.001.
Angel, S., Parent, J., Civco, D. L., Blei, A., & Potere, D. (2011). The dimensions of global
urban expansion: Estimates and projections for all countries, 2000–2050. Progress in
Planning, 75(2), 53–107. https://doi.org/10.1016/j.progress.2011.04.001.
Betru, T., Tolera, M., Sahle, K., & Kassa, H. (2019). Trends and drivers of land use/land
cover change in Western Ethiopia. Applied Geography, 104, 83–93. https://doi.org/
10.1016/j.apgeog.2019.02.007.
Bettencourt, L., & West, G. (2010). A unied theory of urban living. Nature, 467(7318),
912. https://doi.org/10.1038/467912a.
Bhatta, B., Saraswati, S., & Bandyopadhyay, D. (2010). Urban sprawl measurement from
remote sensing data. Applied Geography, 30(4), 731–740. https://doi.org/10.1016/j.
apgeog.2010.02.002.
Bramley, G., Dempsey, N., Power, S., Brown, C., & Watkins, D. (2009). Social
sustainability and urban form: Evidence from ve British cities. Environment &
Planning A, 41(9), 2125–2142. https://doi.org/10.1068/a4184.
Bun, M. J., & Harrison, T. D. (2018). OLS and IV estimation of regression models
including endogenous interaction terms. Econometric Reviews, 1–14. https://doi.org/
10.1080/07474938.2018.1427486.
Burton, E. (2000). The compact city: Just or just compact? A preliminary analysis. Urban
Studies, 37(11), 1969–2006. https://doi.org/10.1080/00420980050162184.
Burton, E. (2002). Measuring urban compactness in UK towns and cities. Environment and
Planning B: Planning and Design, 29(2), 219–250. https://doi.org/10.1068/b2713.
Buxton, M., & Tieman, G. (2005). Patterns of urban consolidation in Melbourne:
Planning policy and the growth of medium density housing. Urban Policy and
Research, 23(2), 137–157. https://doi.org/10.1080/08111470500135086.
Calzada, L., Meave, J. A., Bonl, C., & Figueroa, F. (2018). Lands at risk: Land use/land
cover change in two contrasting tropical dry regions of Mexico. Applied Geography,
99, 22–30. https://doi.org/10.1016/j.apgeog.2018.07.021.
de Espindola, G. M., da Costa Carneiro, E. L. N., & Façanha, A. C. (2017). Four decades of
urban sprawl and population growth in Teresina, Brazil. Applied Geography, 79,
73–83. https://doi.org/10.1016/j.apgeog.2016.12.018.
Ewing, R. (1997). Is los angeles-style sprawl desirable? Journal of the American Planning
Association, 63(1), 107–126. https://doi.org/10.1080/01944369708975728.
Grifths, P., Hostert, P., Gruebner, O., & van der Linden, S. (2010). Mapping megacity
growth with multi-sensor data. Remote Sensing of Environment, 114(2), 426–439.
https://doi.org/10.1016/j.rse.2009.09.012.
Güneralp, B., Zhou, Y., Ürge-Vorsatz, D., Gupta, M., Yu, S., Patel, P. L., & Seto, K. C.
(2017). Global scenarios of urban density and its impacts on building energy use
through 2050. Proceedings of the National Academy of Sciences, 114(34), 8945–8950.
https://doi.org/10.1073/pnas.1606035114.
Haase, D., Kabisch, N., & Haase, A. (2013). Endless urban growth? On the mismatch of
population, household and urban land area growth and its effects on the urban
debate. PLoS One, 8(6), Article e66531. https://doi.org/10.1371/journal.
pone.0066531.
Henríquez, C., Az´
ocar, G., & Romero, H. (2006). Monitoring and modeling the urban
growth of two mid-sized Chilean cities. Habitat International, 30(4), 945–964.
https://doi.org/10.1016/j.habitatint.2005.05.002.
Kohli, D., Sliuzas, R., & Stein, A. (2016). Urban slum detection using texture and spatial
metrics derived from satellite imagery. Journal of Spatial Science, 61(2), 405–426.
https://doi.org/10.1080/14498596.2016.1138247.
M.R. Rahnama et al.
Applied Geography 124 (2020) 102318
9
Kotharkar, R., Bahadure, P., & Sarda, N. (2014). Measuring compact urban form: A case
of Nagpur city, India. Sustainability, 6(7), 4246–4272. https://doi.org/10.3390/
su6074246.
Kumar, A. (2017). Analysing urban sprawl and land consumption patterns in major
capital cities in the Himalayan region using geoinformatics. Applied Geography, 89,
112–123. https://doi.org/10.1016/j.apgeog.2017.10.010.
Lal, K., Kumar, D., & Kumar, A. (2017). Spatio-temporal landscape modeling of urban
growth patterns in Dhanbad Urban Agglomeration, India using geoinformatics
techniques. The Egyptian Journal of Remote Sensing and Space Science, 20(1), 91–102.
https://doi.org/10.1016/j.ejrs.2017.01.003.
Lee, J., Kurisu, K., An, K., & Hanaki, K. (2014). Development of the compact city index
and its application to Japanese cities. Urban Studies, 52(6), 1054–1070. https://doi.
org/10.1177/0042098014536786.
Liao, F. H., & Wei, Y. D. (2014). Modeling determinants of urban growth in Dongguan,
China: A spatial logistic approach. Stochastic Environmental Research and Risk
Assessment, 28(4), 801–816. https://doi.org/10.1007/s00477-012-0620y.
Li, C., Li, J., & Wu, J. (2018). What drives urban growth in China? A multi-scale
comparative analysis. Applied Geography, 98, 43–51. https://doi.org/10.1016/j.
apgeog.2018.07.002.
Lim, H. K., & Kain, J. H. (2016). Compact cities are complex, intense and diverse but: Can
we design such emergent urban properties? Urban Planning, 1(1), 95–113. https://
doi.org/10.17645/up.v1i1.535.
Liu, Y., He, Q., Tan, R., Liu, Y., & Yin, C. (2016). Modeling different urban growth
patterns based on the evolution of urban form: A case study from Huangpi, Central
China. Applied Geography, 66, 109–118. https://doi.org/10.1016/j.
apgeog.2015.11.012.
Liu, Y., Zhang, X., Kong, X., Wang, R., & Chen, L. (2018). Identifying the relationship
between urban land expansion and human activities in the Yangtze River Economic
Belt, China. Applied Geography, 94, 163–177. https://doi.org/10.1016/j.
apgeog.2018.03.016.
McCormack, G. R., & Shiell, A. (2011). In search of causality: A systematic review of the
relationship between the built environment and physical activity among adults.
International Journal of Behavioral Nutrition and Physical Activity, 8(1), 125. https://
doi.org/10.1186/1479-5868-8-125.
Mendiratta, P., & Gedam, S. (2018). Assessment of urban growth dynamics in Mumbai
Metropolitan Region, India using object-based image analysis for medium-resolution
data. Applied Geography, 98, 110–120. https://doi.org/10.1016/j.
apgeog.2018.05.017.
Mouratidis, K. (2017). Is compact city livable? The impact of compact versus sprawled
neighbourhoods on neighbourhood satisfaction. Urban Studies, 55(11), 2408–2430.
https://doi.org/10.1177/0042098017729109.
Naserikia, M., Asadi Shamsabadi, E., Raeian, M., & Leal Filho, W. (2019). The urban
heat island in an urban context: A case study of Mashhad, Iran. International Journal
of Environmental Research and Public Health, 16(3), 313. https://doi.org/10.3390/
ijerph16030313.
OECD. (2012). Compact city policies: A comparative assessment. OECD green growth studies.
OECD Publishing.
Ogle, J., Delparte, D., & Sanger, H. (2017). Quantifying the sustainability of urban
growth and form through time: An algorithmic analysis of a city’s development.
Applied Geography, 88, 1–14. https://doi.org/10.1016/j.apgeog.2017.08.016.
Paulsen, K. (2014). Geography, policy or market? New evidence on the measurement and
causes of sprawl (and inll) in US metropolitan regions. Urban Studies, 51(12),
2629–2645. https://doi.org/10.1177/0042098013512874.
Pearsall, H. (2017). Staying cool in the compact city: Vacant land and urban heating in
Philadelphia, Pennsylvania. Applied Geography, 79, 84–92. https://doi.org/10.1016/
j.apgeog.2016.12.010.
Rahnama, M. R., Wyatt, R., & Heydari, A. (2015). What happened from 2001 to 2011 in
Melbourne? Compactness versus sprawl. Sustainable Cities and Society, 19, 109–120.
https://doi.org/10.1016/j.scs.2015.07.001.
Schindler, M., & Caruso, G. (2014). Urban compactness and the trade-off between air
pollution emission and exposure: Lessons from a spatially explicit theoretical model.
Computers, Environment and Urban Systems, 45, 13–23. https://doi.org/10.1016/j.
compenvurbsys.2014.01.004.
Sharaf, S. A., Serra, P., & Saurí, D. (2018). A district and sector land-use and landscape
analysis of urban sprawl in Al Ain municipality (United Arab Emirates): Just a quick
conversion from sand to a built-up environment? Applied Geography, 95, 88–100.
https://doi.org/10.1016/j.apgeog.2018.04.008.
Shirowzhan, S., & Sepasgozar, S. M. (2019). Spatial analysis using temporal point Clouds
in advanced GIS: Methods for ground elevation extraction in slant areas and building
classications. ISPRS International Journal of Geo-Information, 8(3), 120. https://doi.
org/10.3390/ijgi8030120.
Shum, K., & Watanabe, C. (2017). From compact city to smart city: A sustainability
science synergy perspective. Journal of Environmental Science and Engineering, 4,
200–208. https://doi.org/10.17265/2162-5298/2017.04.004.
Theobald, D. M. (2005). Landscape patterns of exurban growth in the USA from 1980 to
2020. Ecology and Society, 10(1).
Westerink, J., Haase, D., Bauer, A., Ravetz, J., Jarrige, F., & Aalbers, C. B. (2013). Dealing
with sustainability trade-offs of the compact city in peri-urban planning across
European city regions. European Planning Studies, 21(4), 473–497. https://doi.org/
10.1080/09654313.2012.722927.
Wolff, M., Haase, D., & Haase, A. (2018). Compact or spread? A quantitative spatial
model of urban areas in europe since 1990. PLoS One, 13(2), Article e0192326.
https://doi.org/10.1371/journal.pone.0192326.
Wolff, M., Haase, A., Haase, D., & Kabisch, N. (2016). The impact of urban regrowth on
the built environment. Urban Studies, 54(12), 2683–2700. https://doi.org/10.1177/
0042098016658231.
Zha, Y., Gao, J., & Ni, S. (2003). Use of normalized difference built-up index in
automatically mapping urban areas from TM imagery. International Journal of Remote
Sensing, 24(3), 583–594. https://doi.org/10.1080/01431160304987.
Zhang, Z., Su, S., Xiao, R., Jiang, D., & Wu, J. (2013). Identifying determinants of urban
growth from a multi-scale perspective: A case study of the urban agglomeration
around Hangzhou bay, China. Applied Geography, 45, 193–202. https://doi.org/
10.1016/j.apgeog.2013.09.013.
Zheng, H. W., Shen, G. Q., Wang, H., & Hong, J. (2015). Simulating land use change in
urban renewal areas: A case study in Hong Kong. Habitat International, 46, 23–34.
https://doi.org/10.1016/j.habitatint.2014.10.008.
M.R. Rahnama et al.